BackgroundThe relative importance of HDL2 and HDL3 cholesterol as risk factors for ischemic heart disease (IHD) is still uncertain. Their associations with the incidence of IHD in the Caerphilly and Speedwell prospective studies are described.Methods and ResultsThe two studies have a common core protocol and are based on a total of 4860 middle-aged men from the general population. The first follow-up was at a nearly constant interval of 5.1 years in Caerphilly and 3.2 years in Speedwell: 251 major IHD events had occurred. Lipid levels were measured on fasting samples. Different laboratories were used by the two studies. Each laboratory used ultracentrifugation to separate HDL2and HDL3. Both subfractions were inversely associated with risk of IHD. Standardized relative odds of developing major IHD were 0.95 (95% confidence interval [CI], 0.80 to 1.14) for HDL2 cholesterol and 0.83 (95% CI, 0.68 to 1.00) for HDL3 cholesterol in Caerphilly and 0.76 (95% CI, 0.57 to 1.01) for HDL2 and 0.64 (95% CI, 0.49 to 0.83) for HDL3 in Speedwell. The association with incident IHD appeared to be stronger for HDL3in both areas. No linear combination of the two subfractions was a better predictor of IHD than total HDL cholesterol alone.ConclusionsIn British men, both HDL2and HDL3cholesterol are inversely associated with the incidence of IHD. However, the prediction of the risk of IHD from total HDL cholesterol alone could not be improved upon by measurement of the two HDL subfractions. The relative value of the two HDL subfractions as predictors of risk is still unresolved. The uncertainty may be due, at least in part, to problems associated with their measurement.